Hi Karl,

On Jan 6, 2011, at 10:49 AM, Karl Wright wrote:

> The whole question of ease-of-use is what drove this packaging
> arrangement.  I was told it was unacceptable to not have a working
> example out of the box that could be executed in a single line.  Build
> and execution Instructions which involve obtaining a couple of dozen
> jars from other places do not fulfill this criterion.

It's unacceptable to not release software according to Apache guidelines. 
There's some flexibility in those guidelines (whether to include a binary 
release or not, whether to include jar files in a distro or use Maven, etc.), 
and then there's not (must include a source release; must have a KEYS file; 
etc.etc.)

Including a working out of the box example might be something that ManifoldCF 
as a community deems absolutely required for a ManifoldCF release, but it's not 
absolutely required for an Apache release. Of course, not doing it has its own 
implications (users won't care about the software; will find it too difficult), 
but that's another subject.

> 
> The three copies of the dependent jars occur because of the following:
> 
> - There is one copy of the jar that is used by the build
> - There are two distinct execution environments, one single-process,
> and one multi-process, that are built
> - Each execution environment has its own subtree that it executes from

Are all of these Jars simply copies of an original Jar, or are they separately 
licensed?

> 
> If the built environments are no longer distributed, then there will
> be one copy of each dependent jar included.  I'm leaning towards just
> having this minimum distribution since size is apparently a huge issue
> here.

It's a huge issue everywhere. Your release will be mirrored around the world 
using Apache's mirroring system. Beyond that it will be likely replicated N 
times at M companies who are using it. Size *is* a big issue, not just *here*.


>  I still want to know if the source distribution should have the
> forrest-built docs or not, though, or whether it should be up to the
> user to build their own docs using Forrest themselves.  I would prefer
> the former because Forrest is somewhat idiosyncratic, but you guys are
> the bosses.

Eh, either way is fine with me, and I don't think anyone here should legislate 
on this. It should be a ManifoldCF community decision IMHO.

Cheers,
Chris


> 
> Karl
> 
> On Thu, Jan 6, 2011 at 1:39 PM, sebb <seb...@gmail.com> wrote:
>> On 6 January 2011 18:23, Karl Wright <daddy...@gmail.com> wrote:
>>>>> Very well; we will discontinue all binary distributions.
>>>> 
>>>> That's not what I said.
>>>> 
>>>> You can have a binary distribution if you wish, but there must be a
>>>> source distribution.
>>>> 
>>> 
>>> As I said before, it makes no sense to distribute ManifoldCF binaries
>>> without complete sources.  So we could (I suppose) have a source
>>> distribution AND a source+binary distribution.
>> 
>> That would be fine.
>> 
>>>  But we could not simply have a binary distribution and a source 
>>> distribution.
>> 
>> That would also work, but would require binary users to download both 
>> archives.
>> 
>> ==
>> 
>> On a separate matter, I question whether the current packaging is optimal.
>> There appear to be 3 copies of every jar in the binary zip file - no
>> wonder the file is so large!
>> 
>> Also, many of the included jars are commonly used elsewhere, so the
>> consumer may well already have a copy.
>> 
>> Generally, the binary jar consists of the compiled source files only.
>> 
>> Some projects provide additional bundles which include all the
>> required dependencies; that might be the way to go here.
>> 
>>> Karl
>>> 
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
>>> For additional commands, e-mail: general-h...@incubator.apache.org
>>> 
>>> 
>> 
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
>> For additional commands, e-mail: general-h...@incubator.apache.org
>> 
>> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
> 


++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Email: chris.a.mattm...@nasa.gov
WWW:   http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++


---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Reply via email to